Efficient potential of mean force calculation from multiscale simulations: solute insertion in a lipid membrane

TL;DR

This paper evaluates a multiscale simulation method for calculating potentials of mean force during solute insertion into lipid membranes, demonstrating it is accurate and significantly faster than traditional atomistic simulations.

Contribution

It systematically assesses the efficiency and accuracy of a multiscale backmapping approach for membrane solute insertion, highlighting the importance of configurational overlap.

Findings

The method achieves about tenfold reduction in computational time.

Accurate results depend on good configurational overlap between models.

The approach is validated for DMPC and DOPC membranes.

Abstract

The determination of potentials of mean force for solute insertion in a membrane by means of all-atom molecular dynamics simulations is often hampered by sampling issues. A multiscale approach to conformational sampling was recently proposed by Bereau and Kremer (2016). It aims at accelerating the sampling of the atomistic conformational space by means of a systematic backmapping of coarse-grained snapshots. In this work, we first analyze the efficiency of this method by comparing its predictions for propanol insertion into a 1,2-Dimyristoyl-sn-glycero-3-phosphocholine membrane (DMPC) against reference atomistic simulations. The method is found to provide accurate results with a gain of one order of magnitude in computational time. We then investigate the role of the coarse-grained representation in affecting the reliability of the method in the case of a 1,2-Dioleoyl-sn-glycero-3-phosphocholine membrane (DOPC). We find that the accuracy of the results is tightly connected to the presence a good configurational overlap between the coarse-grained and atomistic models---a general requirement when developing multiscale simulation methods.